Afya Pamoja (Health Together)

Within primary healthcare facilities in Tanzania there are inadequate channels for patients to provide feedback on the services and care they receive. This means that there is a lack of available data providing localized real-time insights on primary care facility performance sourced directly from those receiving care. The typical facility receives just two pieces of patient feedback per month - typically via paper and pen notes put into suggestion boxes. 60% of patients say that existing feedback mechanisms are inadequate. As a result, patient voices are not sufficiently heard within healthcare decision making. 

Robust research shows that patient feedback services can have a significant impact on healthcare outcomes, such as reducing maternal mortality by 38%. Insufficient patient feedback reduces the ability of healthcare workers and public healthcare managers to understand the real-time needs and challenges of patients on a local level. This limits their ability to provide targeted oversight and support to healthcare workers or patients (e.g. identifying abusive healthcare workers or training requirements for quality improvement), and reduces their ability to understand resource needs (e.g. identifying where antenatal folic acid supplements are out of stock), advocate for more resources effectively and then allocate them efficiently to facilities (e.g. prioritizing the distribution of limited resources to facilities based on reliable and comparable real-time indicators). 

Solving this feedback problem ensures citizens’ voices can be heard, drives improved top-down and bottom-up social accountability, and supports performance management and quality improvement within facilities. This also results in improved trust between communities and healthcare workers leading to higher utilization of available health services and improved availability of quality healthcare services.

The scale of this problem is large. In Tanzania, 6,000 public primary healthcare facilities provide essential healthcare services to 51 million people. Women and children are particularly reliant on these services for antenatal and early childhood care. These facilities also disproportionately serve low-income and rural communities who rely on public facilities for their healthcare needs. Evidence shows that this service can drive impact across LMICs and our intention is to expand into more countries.

Our innovation is a digital patient feedback service. Using free SMS surveys patients provide feedback, through their own phones, on the quality of services and experience of care they received at a healthcare facility. This data is analyzed and provided to healthcare workers and government officials in the form of easy-to-use alerts, dashboards, and quality reports. The service provides healthcare staff with large volumes of real-time, facility specific and actionable insights on the quality of services provided to drive responsive and accountable decision making. This data source can be used as a proxy to inform performance improvement. The service can be used at all levels of the healthcare system, in communities to catalyse social dialogues during village nutrition days, in facilities for Quality Improvement meetings and during Facility Governing Committees and by local government public health authorities during Supportive Supervision and during annual district-level planning processes. Lastly, updates are provided back to patients to close the feedback loop and build trust. 

Insights are provided on the availability of services (such as those required during antenatal care visits) and the quality of services (such as family planning guidance). Healthcare workers can then identify and address the needs of patients in a systematic manner through supportive performance management tools. Government health officials can also identify service challenges (such as healthcare worker attendance) and gaps and so divert support (such as audits) and resources to these issues. In designing our service we conducted a journey mapping exercise to understand the actions of patients, healthcare workers and government health officials and we held extensive interviews with all stakeholders to determine how data should be collected from patients and presented and disseminated to healthcare workers and local officials.  

Compared to existing feedback mechanisms (such as hand-written notes put into suggestion boxes) a digital service addresses barriers for patients and government. For patients it resolves the key fear around anonymity and reduces the literacy barrier to providing feedback. For the government, the innovation increases the volume of feedback collected at a low cost and without need for additional burden to healthcare workers or officials. It also enables it to be analysed in real-time in a systematic way and directed to multiple levels of the healthcare system. The solution includes a level of flexibility such that healthcare managers can collect data on specific issues as and when they need, receiving specific information on their highest priority issues. Lastly, to ensure adoption and sustainability, our data is integrated into existing government health information systems (DHIS2) and insights directed through existing performance management processes (such as quarterly facility supportive supervisions and monthly facility quality improvement meetings).

We believe this solution has the potential to be the most cost-effective life saving intervention in global health. Rigorous academic research in this field shows that improving patient feedback processes can have dramatic impacts on healthcare outcomes such as reducing child mortality by 38%. This solution builds on the benefits of traditional social accountability interventions (such as social dialogues and community score cards) by utilizing low-cost, digital technologies which are scalable and ensure data can be used at multiple levels of the healthcare system. Our approach blends bottom-up citizen feedback with top-down accountability by influencing government decision making, which helps drive systems change impact through the healthcare system. 

In the pilot phase of our work, we are focusing on maternal, newborn and early childhood healthcare. The key beneficiaries of our intervention will be pregnant women, new mothers and young children. As a nationwide service it would impact those in both rural and urban settings. Public primary care clinics are disproportionately used by low-income houses making the impact skewed towards these households. 

These populations are currently underserved , in Tanzania there are 11,000 maternal and 103,000 under-5 deaths per year. Moreover, the rate of maternal mortality is 50% higher than in neighboring Kenya and Uganda. In the long-term our key outcome metrics will be maternal mortality, infant mortality and under-5 mortality.

Replicating results from previous studies of patient feedback services, could reduce these metrics by 38%. At a national scale in Tanzania would result in 40,000 fewer under-5 deaths per year. 

Dr. Helga Mutasingwa, Chief Program Officer is a Tanzanian doctor and public healthcare professional with 10 years of experience. She managed ICAP’s HIV-patient retention program in the Dar es Salaam Region, visiting public healthcare facilities, patients, facility staff and local government health monitoring teams daily. She has a deep, personal understanding of the challenges faced by the Tanzanian healthcare system due to lack of feedback and has a strong network throughout government. She is an Echoing Green Fellow in recognition of her co-founding of Afya Pamoja.

Marcos Mzeru, Chief Systems Officer, has spent 15 years designing digital health IT systems within the Ministry of Health for the government of Tanzania. He has previously built a patient feedback system.  He has both a detailed knowledge of the health information systems within the government and extensive relationships throughout government.

Simon DeBere, Team Lead, has spent 6 years working on service delivery interventions in East Africa including roles within social enterprises, government and CivicTech organizations.

We have built a team of 8 program staff who are a mixture of doctors, nurses and public health professionals who spend their time implementing our program with communities in four regions of Tanzania. They each spend 80% of their time working in healthcare facilities and local government health offices engaging with patients, healthcare workers and government officials to understand how our service can be improved.

During the design phase of our program, we interviewed 150 patients, healthcare workers, government officials and academics to understand the local challenges related to patient feedback and available performance management data. Following this we established a steering committee with the Government of Tanzania with representatives from across the Ministry of Health (including the Director of IT) and Ministry of Local government (including the Director of Health). This Steering committing provides quarterly feedback on viability and progress of our solution and maintains overall control of its strategic direction. 

The technology platform for our service is already established and we are live in 50 facilities in Dodoma Region with hundreds of pregnant mothers registering to the platform every week, providing a steady flow of feedback data. As a result of our team’s work to get to this stage we have shown high levels of commitment to solving this problem.

We recently launched our pilot in 50 healthcare facilities, and within the first month we have received feedback from 1,000 clients and collected over 10,000 data points. Over the coming year we will be testing a range of program design and technology innovations in order to determine how we can drive maximum impact from this data and service. We believe this challenge can help us do that. There are three key barriers we hope to overcome with support of this challenge:

• Firstly, we need to improve our understanding of how citizen feedback data can be aggregated with other complementary datasets in order to understand primary healthcare delivery and quality. Through this challenge we hope to learn for our peer organistions about other unconventional data sources on quality and therefore better fit citizen feedback data into this analysis. We have a range of hypotheses about where citizens may and may not be well placed to supplement existing data and therefore refining this understanding is vital.

• The second is better understanding how to effectively utilize citizen feedback with different healthcare settings from facility level settings to national government level policy environments. We are currently embarking on a 12-month period of intensive testing and experimentation to grow our understanding in this space and hearing from others doing data innovation work would be very useful.

• Finally, the challenge would provide us with improved access to a network of potential funding partners who are interested in our work and may wish to support our data use innovation workstream.

Our solution is innovative for the following reasons:

• We collect high volumes of real-time data on performance: Existing analogue feedback mechanisms collect low volumes of data (2 data points per month per facility, on average) and other health information systems report data only monthly.

• We collect data which provides better insights on the quality of care: Relative to existing data sources we can collect indicators which better proxy for the quality of care being provided rather than just the volume of healthcare services delivered. As result, our data is more useful for identifying areas for quality improvement. 

• We collect data directly from patients: This not only elevates the voices of patients who directly receive care but also minimizes the risk of certain biases in collection found in other data sources. Using patients as the data source also minimizes any additional burden on healthcare worker or the need for additional data collection tools, which may be expensive and unscalable.

• We use widely accessible technologies which require limited connectivity: SMS enabled phones are owned by 85% of citizens in Tanzania and service coverage is found across most of the country. Using the service is free to citizens.

• We integrate our data into the government’s existing health information systems: The most common portal for healthcare workers and government officials accessing health data is DHIS2 and our data integrates directly into this system so that it can be used in a sustainable manner with strong government “buy-in”

• We have developed a range of decision-making tools to support the use of our data: This includes dashboards, quality reports and alerts based on the analyzed data. These have been integrated at different levels of the healthcare system including with quality improvement and governing meetings in facilities and supportive supervisions and resource planning meetings within district health authorities.

• We can flexibly adjust the data collected based on specific or evolving needs: The digital platform means that we can tailor the data we collect from different types of patients or based on different types of care they receive. We are also able to update the data collected periodically and in line with government needs (e.g. disease outbreaks). We believe that such a flexible system could be catalytic in how patient feedback.  

• We can integrate with other complementary digital health programs: Our technology has open APIs and uses commonly-used open-source technology, enabling us to integrate with other services such as outbound SMS messaging to pregnant mothers.

Next year:

Outcomes:

1. Increase utilization of at least one service (ANC / LND / PNC) by 20% 
2. Increase availability of at least one service (testing, supplements, medications, supplies, presence of qualified staff) by 20%
3. Increase or improve at least one aspect of service quality (hygiene, staff friendliness, short waiting times) by 20%

Outputs

1. 30% of patients within our initial target population of patients receiving reproductive and child health care complete feedback surveys
2. 30% increase in the use of client feedback in at least one of the following settings: local government supportive supervisions at facilities, facility quality improvement meetings, facility governing committee meeting
3. 10% of community members attend a community social dialogue each quarter where patient feedback data is discussed

Scale

1. We will operate our service in 500 primary care facilities in 3 regions in Tanzania and over 1 million citizens will have provided feedback via our platform
2. We will have launched a pilot of our service in one additional country in Africa 

Next five years: 

Impact

1. Reduced maternal and child mortality by 25% 

Outcomes

1. Increase utilisation of key services (ANC / LND / PNC) by 30% 
2. Increase availability of key services (testing, supplements, medications, supplies, presence of qualified staff) by 30%
3. Increase service quality (hygiene, staff friendliness, short waiting times) by 30%

Outputs

1. 30% of patients complete feedback surveys
2. 50% increase in use of client feedback in the following settings: supportive supervisions, facility quality improvement meetings, facility governing committee meeting
3. 10% of community members attend a community social dialogue each quarter

Scale

1. We will operate a nationally scaled, sustainably-funded digital feedback service in all 6,000 of Tanzania’s public primary care facilities and the service will be fully integrated into government technology systems 
2. We will have expanded the service into 3 additional countries across Africa, launching pilots in each of them and be on course to expand nationally

Using the traditional logic framework, we will measure indicators across four areas:

Impact (although it’s too early to see changes in the indicators we have set-up a system to track changes in these metrics in the next 2 years)

1. Maternal, infant and under-5 mortality indicators 

Outcomes (we already have systems in place measuring these indicators)

1. % of women attending 4+ ANC visits, a PNC visit and have an institutional delivery (initially out service is focused on reproductive and child healthcare)
2. Client reported availability of services 
3. Client reported satisfaction with hygiene, staff behaviour and waiting times

Outputs (we already have systems in place measuring these indicators)

1. % of pregnant women registering for our service and completing the feedback survey
2. % of facilities conducting Supportive Supervision, Quality Improvement Meeting and Facility Governing Committees meeting use client feedback data from our platform

3. # of community members attending a social dialogue in the community where client feedback data from our platform is discussed

Activities (we already have systems in place measuring these indicators)

1. # of healthcare workers and community healthcare workers trained on our platform and marketing 
2. # of facility in-charges, facility governing committees, district health teams and facility quality improvement teams provided with client feedback data from our platform
3. # of social dialogues held in the community 

From reviewing over 10 rigorous studies of accountability mechanisms in healthcare, our theory of change centres on two sets of activities. Firstly, it’s possible to collect and analyze large volumes of patient feedback - demonstrated in Kenya and Tanzania (Okeke 2019, UNICEF 2020). Previous pilots (Grossman 2017, UNICEF 2020) have shown that mobile patient feedback systems result in patients providing up to 200 times more feedback, demonstrating a willingness to adopt such services. We build on this work by collecting data from patients in a structured, more actionable format. Secondly, sharing targeted patient-centred insights with healthcare workers and officials can drive behaviour change, increase accountability and improve decision making (e.g. planning, resourcing, performance management).

These activities drive changes in three types of outcomes. Firstly, increasing trust in services leads to increased service utilisation (Mselle 2019). Improved trust is achieved by allowing patients to provide feedback on the services and following up on this feedback. Evidence from similar services suggests that this can increase utilization by 20% (e.g. antenatal care visits) (Gullo 2016). Secondly, this service can increase service availability. This is achieved by better understanding patient needs and increasing oversight of individual healthcare workers and facilities. Studies (Callen 2020) find that providing data to managers increases facility oversight visits by 104% and doctor attendance by 75%. This results in improved resource allocations (e.g. medicine and staff allocations) and improved staff behaviours. Evidence from similar services suggests that it can increase availability of essential medicines by 41% (Blake 2016).Thirdly, this service can increase service quality with studies suggesting that patient feedback can improve facility hygiene, staff friendliness, patient waiting times (UNICEF 2020).

The impact of these changes in outcomes are on maternal, infant and under-5 mortality. Seminal work found that similar accountability mechanisms can reduce under-5 mortality by 33% (Bjorkman 2009). A more recent RCT of over 250 health clinics in Sierra Leone found that patient feedback services reduced under-5 mortality by 38% (Christensen 2021). They also found that these effects persist during disease epidemics resulting in case reporting increasing by 62%. 

Afya Pamoja’s solution builds on the benefits of traditional social accountability interventions (such as social dialogues and community score cards) by utilizing low-cost, digital technologies which are scalable. Unlike recent work (Raffler 2020, Arkedis 2021) which finds limited impact from just bottom-up community feedback mechanisms, our approach blends bottom-up feedback with top-down accountability by influencing government decision making. By supporting decision makers, our work builds on recent work (Kosack 2021) which suggests accountability programs involving officials are more effective than those which rely on social accountability alone.

Our core value is to use existing, well established mobile technologies and deploy them into a new setting. We create value by deeply integrating these technologies within government institutions.

Afya Pamoja is an accessible digital solution. We rely on using widely accessible and relatively simple mobile technologies (e.g. SMS and USSD) to connect patients with medical professionals and public health managers. With mobile penetration now over 85% in Tanzania, now is a unique time when we can connect with nearly the entire population in a low-cost and scalable manner. However, given that smartphone penetration is just 25% and largely concentrated in urban areas, we will focus on feature phone technologies to give a voice to the widest set of citizens possible. Our goal is not to be driven by technology but rather to be driven by the needs of citizens.

Technology enables direct citizen engagement: Our ideas are rooted in the idea that citizens understand their experiences best and want to share views on public services. Afya Pamoja is the only direct patient feedback service in Tanzania. Although other feedback services exist - for example through community healthcare workers - none leverage the high mobile penetration and directly target the patient. The underlying technology for mobile surveys has been thoroughly tested across East Africa and therefore our innovation is connecting this communication technology to effective decision-making tools for public health managers. The challenge does not primarily lie in collecting the data but rather in our ability to leverage the data collected from the technology to derive meaningful and action orientated insights. This end-to-end solution is our innovation.

Our data is collected directly from patients by SMS feedback surveys which they complete following a visit to a primary healthcare facility. They have firsthand experience of the services they receive and then immediately following this interaction register to our service, supported by community healthcare workers where necessary. Shortly after registering patients receive a series of survey questions. Summaries of the feedback and resulting actions taken are shared back with community members to close the feedback loop and build trust in the platform.

Patients are motivated to register and respond to surveys for three reasons. Firstly, they are provided with clear marketing on how their feedback will be used and why it is being collected, thus assisting them in understanding their role in improving the services they receive. Secondly, they are encouraged by community healthcare workers (who are financially incentivized to support the platform to collect data) about why and how they should use the platform. Lastly, the platform contains additional features which directly benefit the patient. For example, appointment reminders, healthcare information, advice, and nudges. These features make the information sharing 2-way and reinforce the value of the platform to patients.

From our founding team's inception this has been a core and on-going conversation within our work. Our founding team comes from a diversity of backgrounds and we believe this is core to our success. We believe open conversation among all team members accompanied by a willingness to take a critical view about DEI issues within our organization is vital. 

Our objectives are primarily two fold, firstly to recruit and entirely staff from the regions we work in. Our goal is to build an East African led organization. This means actively searching for top talent in our markets, fairly assessing it and then investing in it. To date we have delivered on this goal and have hired 15 Tanzanians into our organization. We have agreed to heavily scrutinize any non-East African hire we may be interested in hiring in the future. 

Secondly, we are focused on building a transparent and collaborative culture among both our leadership team and at all levels of the organization. We want a diverse set of perspectives to have oversight of most decisions and be empowered to share their views freely. We recognize that this is more than just a process and requires on-going investment to ensure this culture is established and maintained. We discuss this topic in monthly team meetings and dedicate part of our quarterly team retreat to this topic.

Our service is free-to-use for patients and healthcare facilities. We operate as a non-profit and receiving funding from a range of sources. 

Government cost-sharing: Within 3 years we expect to achieve government cost-sharing arrangements with local government authorities covering a portion of our operating costs, specifically those related to fees incurred in working with facility staff who are already on the government payroll. We are encouraged that a similar organisation, Jacaranda Health, in Kenya, has been able to achieve such a funding arrangement with the Kenyan government.

SMS fee subsidies: Ministry of Health has access to a zero-rated SMS short-code for digital health services already agreed with Tanzania’s four major telecommunication companies. Once we have piloted our program and proven our ability to deliver impact we will negotiate access to this short-code and receive a zero-rating. These conversations have begun already and will begin these tests before the end of the year. 

Revenue-generation: Longer-term we want to develop a series of complementary revenue-generating services. Given we have privileged data from and access to reach a segment of care-seeking individuals we believe that we can sell health insurance, medical expense loans and a range of services to pharmaceutical companies, to subsidise our operating costs. 

Grant funding: Initially our service will be funded through philanthropic capital from global health and civic engagement funders who are motivated by high-impact, low-cost, scalable ventures. Initially this money will be raised from philanthropic capital and will be sufficient for us to test our product and expand to a scale at which we can undertake rigorous testing on the impact of our service, by partnering with academics to run a Randomised Control Test. By proving our impact through the results of this study we will be able to crowd-in funding from funders such as Global Innovation Fund, USAID DIV and FID, who provide multi-year large-scale grants for programs with robust impact evidence.  

Overall funding share: At scale, across all 6,000 facilities in Tanzania, our service will cost $6m p.a. to run and within 5 years we expect to cover 40% of this cost through SMS fee subsidies from telecommunications companies, 20% through government cost-sharing arrangements, 20% through long-running grants and 20% through complementary revenue-generating services. 

There are two phases on Afya Pamoja's path to financial sustainability. Phase one runs from the pilot through to the point of having robust evidence for our impact. This phase will last two years and is primarily funded by mobile network subsidies and grant funding. The Tanzanian government has agreed with mobile network operators to zero-rate the SMS costs of this service, which reduces our costs by 60%. 

In phase two, once impact is proven (e.g. through an RCT) and we are scaling to serve all 6,000 facilities, we will reduce our need to raise grant funding by generating income to cover a share of operating costs through a range of sources. 

• Charge or agree cost-sharing arrangements (e.g. paying CHW per diems) with government and their development partners (e.g. bilateral donors). 

• Monetise our data through offering market insights to organisations such as health insurance providers and pharmaceutical companies. 

• Partner with lenders to sell finanical services such as health insurance and loans for medical expenses, charging our partners commission. 

• Using our RCT results we will use innovative outcome funding structures. 

Through this approach we will achieve financial sustainability for a national service within 4 years. 

We have raised $500,000 from a diversified set of funders in the past 12 months from (1) Multilateral organizations in Tanzania (UNICEF), (2) Venture philanthropists (Former Board Chair of Global Innovation Fund), (3) Social entrepreneurship funders (Echoing Green), (4) Development Innovation Funders (Fund for Innovation in Development), (5) Start-up competitions (D-Prize for Development).

We’ve secured a long-term funding partnership with UNICEF for up to $500,000 over the next 2 years. Alongside this we have also begun discussions with USAID to consider funding the roll out of this service in 1500 primary care facilities across Tanzania.

We’ve begun discussions with the government and the mobile network operators to leverage their free to use SMS short code and the government are eager for us to use this short code. We will begin testing our service on this short code by the end of the year.